Efficiency is the ratio of work done to the effort expended. Watts are how we measure the effort expended and dB are how we measure the work done. The only other consideration in this equation is whether the work is being done where and how you want it. One could build a ruthlessly efficient system that was 112dB at 1 watt but did nothing at 30Hz. One could also build a desperately inefficient system that had 60,000W of power that would move more air with the fans in the amps than with the speakers. Buying or renting either system based on the power rating or efficiency alone would be ill advised.
When you hear a system you like, consider what it is that you like about it, what makes you like it. Consider what you might not like about it, too. When you hear a system you don’t like, consider what its doing to make you dislike it. I would imagine that you won’t be able to tell from listening how many watts its using.
We once had a system out at a festival and the promoter wanted to have a BIIIGG system! We had a big system with about 60,000W of amplification behind it. The noise police had us limited to something like 92dB at the property line, so we were holding way back, but how far back, I wondered? I went to the amps, which display the real-time power output in watts and picked the peak numbers displayed from each frequency band and added up the entire system output we were using to achieve what was inevitably over the SPL limit at 100’. The answer: About 1W.
Sensitivity is a specific measurement of efficiency however without a reference to a frequency range it doesn’t tell you enough to make a choice. Efficiency is the ratio of the work done to the effort expended. In this case, Watts are how we measure the effort expended and dB SPL is how we measure the work done.
One could build that desperately inefficient system that had 60,000W of power amps and would move more air with the fans in the amps than with the speakers and it would be easy to sell to people who only ask how many watts.
Alternately, one could build a ruthlessly efficient system that was capable of 115dB at 1 watt. It would be easier to run it off 1 wall outlet but it would be harder to sell to the uninitiated. Given the physics involved, for decent low-range response it would probably also be larger to transport.
To me, the more important consideration in this equation is whether any work is being done where and how you want it. Efficiency isn’t the determining factor for a subjective evaluation of sound quality.
Buying or renting a system based on its power rating alone would be ill advised. Similarly, choosing a system based only on the efficiency rating is also lacking relevant data. Picking a sound system based on how many watts it uses is about as appropriate as picking your girlfriend for how many calories she eats in a day. There are probably more important considerations!
Efficiency can be related to suitability for a particular purpose. For me, the consideration of efficiency isn’t necessarily limited to how many watts a system uses for a given dB SPL. In my view, it could well be expanded to how big a vehicle is required to transport it and how much time it takes to set up and how many people (and sweat) it takes to set it up. In some cases its fair to trade a little electrical energy loss to offset losses elsewhere, especially those losses you feel very personally every time you do a gig!
Power handling is essentially a measurement of thermal dissipation capacity. Higher power handling only really means that you can shove more power in without melting something. Sensitivity expresses the ratio of electrical power input relative to sound power output. To put this into perspective, a loudspeaker that has a sensitivity of 93dB measured with 1 watt at 1m distance is technically only 1% efficient. 96dB 1W@1M is 2% efficient , 99dB 1W@1M is 4% efficient, 102dB 1W@1M is 8% efficient and so on. For every 3dB more output you get for the same input, the electrical to acoustical conversion efficiency doubles. A system that has a sensitivity of 93dB 1W@1M is converting 1% of the electrical energy into acoustical energy and the rest, 99%, is wasted as heat. For a system that has a sensitivity of 105dB 1W@1M, 16% of the electrical energy is being converted into acoustical energy, which is 16x more than the 93dB sensitive system! That’s 1600% more sound for the same watts!
That’s why how many watts are going in is relatively insignificant and what percentage of it that you get out as sound is so significant. Even in the 105dB sensitive system, 84% of the electrical energy is still being lost as heat, so power handling, the capacity to dissipate heat, is still a major concern. A system that is efficient AND has high power handling is most desirable but in the context of loudspeaker drivers, power handling alone is largely an expression of how much energy can be wasted.
What about peak SPL? Industry practice is to quote peak SPL as a number derived from a calculation of the specified sensitivity of the speaker combined with the available amplifier power, or power handling, which is often quoted as a “peak” number. This is a short-term consideration at best and assumes the specifications are accurate, achievable and representative of broad-spectrum performance. In truth they are mostly cherry-picked and chosen to look good on paper. Very rarely can they be achieved. Choosing a system based on a specification sheet’s peak SPL number is like believing everything you read on an internet dating site: A recipe for disappointment.